In K-Means clustering, the first step is usually choose the number of clusters assign centroinds to midpoints calculate the midpoint for each cluster assign each data point to a centroid